

DTM E59. Real Time ML & Generative AI Edge Computing Platform - Varun, NimbleEdge
Jun 7, 2024
Varun Khare, CEO of NimbleEdge and a tech entrepreneur with a rich background in machine learning research, discusses the innovative approach behind NimbleEdge's on-device ML platform. He reveals the gap in the market that led to its creation and how it enables real-time personalization while preserving privacy. Varun shares insights on the success of Dream11, explaining how edge computing optimizes user experience during peak times. He also explores the future of generative AI at the edge, promising transformative potential for mobile applications.
AI Snips
Chapters
Transcript
Episode notes
Varun’s Edge Computing Origins
- Varun Khare describes his journey to edge computing through leading an open source project enabling on-device ML training.
- This project was one of the first to make smartphone-based ML training accessible to the community.
Device Diversity in Edge Computing
- The main challenge in deploying edge computing is handling device diversity across smartphones.
- NimbleEdge provides a unified platform abstracting differences in hardware specs and capabilities.
Why Existing ML Runtimes Fall Short
- Existing ML runtimes do not address key issues like data processing on devices or resource control.
- No unified stack existed to handle device heterogeneity and machine learning lifecycle on mobile.